Diet-responsive transcriptional regulation of insulin in a single neuron controls systemic metabolism

Abstract

Metabolic homeostasis is coordinated through a robust network of signaling pathways acting across all tissues. A key part of this network is insulin-like signaling, which is fundamental for surviving glucose stress. Here, we show that Caenorhabditis elegans fed excess dietary glucose reduce insulin-1 (INS-1) expression specifically in the BAG glutamatergic sensory neurons. We demonstrate that INS-1 expression in the BAG neurons is directly controlled by the transcription factor ETS-5, which is also down-regulated by glucose. We further find that INS-1 acts exclusively from the BAG neurons, and not other INS-1-expressing neurons, to systemically inhibit fat storage via the insulin-like receptor DAF-2. Together, these findings reveal an intertissue regulatory pathway where regulation of insulin expression in a specific neuron controls systemic metabolism in response to excess dietary glucose.

Fig 1. A high glucose diet down-regulates ins-1 promoter activity in the BAG neurons.

(A) A 2.5-kb fragment of the ins-1 promoter drives expression in >10 neurons. (B) Representative images of ins-1p::NLS-GFP expression in E. coli OP50-fed (upper panel) and E. coli OP50 + 40 mM glucose-fed (lower panel) animals (BAG neurons indicated by red arrows). Diet treatment: 24 hours. Scale bar 25 μm. (C) Quantification of nuclear GFP intensity across all observable neurons of the ins-1p::NLS-GFP reporter grown on E. coli OP50 or E. coli OP50 + 40 mM glucose. Diet treatment: 24 hours. Data presented as CTCF % of no glucose, $\overline{x}$ + SEM, n shown within bars, ns, not significant (p > 0.05), *** = p ≤ 0.001, significance assessed by unpaired t test. The underlying numerical data can be found in S1 Data. CTCF, calculated total fluorescence; GFP, green fluorescent protein.

Fig 2. Excess glucose acts directly on the BAG neurons to down-regulate ins-1.

Quantification of nuclear GFP intensity in BAG neurons of the ins-1p::NLS-GFP reporter in (A) OP50 that was heat-killed prior to plating on H₂O or glucose plates, (B) mutant for cilia formation (che-3), (C) mutants for neurotransmission (unc-13), (D) mutants for neuropeptide secretion (unc-31), and (E) animals with BAG-specific RNAi-mediated knock-down of unc-31, grown on E. coli OP50 or E. coli OP50 + 40 mM glucose. (F) Quantification of nuclear GFP intensity in BAG neurons of the ins-1p::NLS-GFP reporter grown on D-glucose, palmitic acid, L-glucose, and D-sorbitol. Diet treatment: 24 hours. Data presented as CTCF % of wild-type on standard diet (H₂O), $\overline{x}$ + SEM, n shown within bars, ns, not significant (p > 0.05), * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, A–D significance assessed by mixed-effects model (2-way ANOVA). E significance assessed by 1-way ANOVA with Dunnet’s correction. The underlying numerical data can be found in S1 Data. CTCF, calculated total fluorescence; GFP, green fluorescent protein; RNA interference.

Fig 3. The BAG-specifying transcription factor ETS-5 regulates ins-1 expression.

(A) The ins-1 promoter contains two adjacent ETS motifs (green boxes). (B) Representative images of ins-1p::NLS-GFP expression in ets-5(tm1734) animals compared to wild-type. Red circles: BAG neurons, blue circle: ASH neurons. Scale bar 20 μm. (C) Quantification of GFP intensity (CTCF) of ins-1p::NLS-GFP in the BAG and ASH neurons in wild-type and ets-5(tm1734) animals. Data presented as $\overline{x}$ + SEM, n shown within bar. Significance assessed by unpaired t test. (D) ETS-5-GFP ChIP-qPCR. Regions assessed: flp-17 promoter and ins-1 promoter containing ETS-sites (green), ins-1 promoter non-ETS-5 site region (gray), and no-antibody control (no-ab). Data are presented as $\overline{x}$ +SEM, n = 4. Significance assessed by ratio paired t test. (E) Representative images of ins-1p::NLS-GFP with both putative ETS sites mutated (ΔΔETS) compared to wild-type. Red circles: BAG neurons. Scale bar 20 μm. (F) Quantification of GFP intensity (CTCF) in BAG relative to wild-type control line #1, with ins-1 promoter mutated at ETS1, ETS2, or both, 2 independent lines assayed, n shown in bar. Significance assessed by 1-way ANOVA with Dunnet’s correction. (G) Representative images of endogenous ETS-5-GFP expression in the BAG neurons (red arrow) of animals fed standard diet (E. coli OP50) (upper panel) and high glucose diet (E. coli OP50 + 40 mM glucose) (lower panel). Diet treatment: 24 hours. Scale bar 20 μm. (H) Quantification of endogenous ETS-5-GFP expression in the BAG neurons of animals fed standard diet (E. coli OP50, gray bar) and high glucose diet (E. coli OP50 + 40 mM glucose, yellow bar), displayed as CTCF values as % of OP50 value. Diet treatment: 24 hours. Data presented as $\overline{x}$ + SEM, n shown in bar. Significance assessed by unpaired t test. For all data: ns, not significant (p > 0.05), * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001. The underlying numerical data can be found in S1 Data. CTCF, calculated total fluorescence; GFP, green fluorescent protein; RNA interference.

Fig 4. INS-1 controls intestinal fat levels.

(A) Quantification of ORO staining of wild-type, ins-1(tm1888) and ins-1(nj32) animals. Data presented as ORO intensity as % wild-type, $\overline{x}$ + SEM, n shown within bar. Significance assessed by 1-way ANOVA (Dunnet’s correction). (B) Representative ORO images of wild-type, ins-1(tm1888) and ins-1(nj32) animals. Scale bar 50 μm. (C) Exploration assay of wild-type, ins-1(tm1888) and ins-1(nj32) animals. Data presented as $\overline{x}$ + SEM, n shown in bar. Statistical significance assessed by 1-way ANOVA with Tukey’s correction for multiple comparisons. (D) Exploration assay of ins-1(nj32) animals on RNAi plates, treated with either empty vector control (gray) or RNAi targeted to pod-2 (red). Data presented as $\overline{x}$ + SEM, n shown in bar. Statistical significance assessed by unpaired t test. (E) Double mutant analysis for ets-5(tm1734) and ins-1(nj32). ORO quantification of ins-1(nj32); ets-5(tm1734) double mutant compared to each single mutant. Data presented as $\overline{x}$ + SEM, n shown within bar. Significance assessed by 1-way ANOVA (Tukey’s correction). For all data: ns, not significant (p > 0.05), * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001. The underlying numerical data can be found in S1 Data. ORO, Oil-Red O; RNAi, RNA interference.

Fig 5. INS-1 regulates intestinal fat levels specifically from the BAG neurons.

(A) Schematic of expression pattern observed for wild-type ins-1 promoter (upper panel) and ins-1 promoter with both ETS site mutated (middle panel), the BAG-specific flp-17 promoter (green neurons, lower panel), and AIY-specific ttx-3 promoter (blue neurons, lower panel). (B) ORO quantification of ins-1(nj32) rescued with wild-type Ex[ins-1p::ins-1cDNA] (green bar), or mutant Ex[ins-1ΔΔETSp::ins-1cDNA] (blue bar), compared to ins-1(nj32) and wild-type. (C) Exploration assay of ins-1(nj32) rescued with wild-type Ex[ins-1p::ins-1cDNA] (green bar), or mutant Ex[ins-1ΔΔETSp::ins-1cDNA] (blue bar), compared to ins-1(nj32) and wild-type. (D) Quantification of ORO staining of ins-1(nj32); Ex[flp-17p::ins-1cDNA] (BAG-specific, green bar) compared to ins-1(nj32) and wild-type. (E) Exploration assay of ins-1(nj32); Ex[flp-17p::ins-1cDNA] (BAG-specific, green bar) compared to ins-1(nj32) and wild-type. (F) Quantification of ORO staining of ins-1(nj32); Ex[ttx-3p::ins-1cDNA] (AIY-specific, blue bar) compared to ins-1(nj32) and wild-type. (G) Exploration assay of ins-1(nj32); Ex[ttx-3p::ins-1cDNA] (AIY-specific, blue bar) compared to ins-1(nj32) and wild-type. All data presented as $\overline{x}$ + SEM, n shown in bar. Significance assessed by 1-way ANOVA (Tukey’s correction). ns, not significant (p > 0.05), * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001. The underlying numerical data can be found in S1 Data. ORO, Oil-Red O.

Fig 6. INS-1 acts from the BAG neurons to control intestinal DAF-2 activity.

(A) ORO quantification of wild-type, ins-1(nj32) and daf-2(e1370) single mutants and the daf-2(e1370); ins-1(nj32) double mutant in the presence or absence of Ex[flp-17p::ins-1cDNA]. Assayed at daf-2(e1370) semi-permissive temperature of 20°C. (B) ORO quantification of daf-2(e1370) single mutants and the daf-2(e1370); ins-1(nj32) double mutant in the presence or absence of Ex[flp-17p::ins-1cDNA]. Shifted to the daf-2(e1370) restrictive temperature 25°C for 24 hours prior to assaying. (C) Schematic diagrams showing where daf-2 is expressed under the rimb-1 (pan-neuronal) and ges-1 (intestinal) promoters. (D) Representative ORO images of wild-type, daf-2(e1370), daf-2(e1370); Ex[rimb-1p::daf-2a] (neuronal daf-2a expression) and daf-2(e1370); Ex[ges-1p::daf-2a] (intestinal daf-2a expression) animal shifted to 25°C for 24 hours prior to staining. Scale bar 50 μm. (E) ORO quantification of wild-type, ins-1(nj32) and daf-2(e1370) single mutants and the daf-2(e1370); ins-1(nj32) double mutant with neuronal (blue bars) and intestinal (orange bars) expression of daf-2a cDNA. All data presented as $\overline{x}$ + SEM, n shown in bar. Significance assessed by 1-way ANOVA (Tukey’s correction). ns, not significant (p > 0.05), * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001. The underlying numerical data can be found in S1 Data. ORO, Oil-Red O.

Fig 7. INS-1 action is DAF-16 dependent and modulates intestinal DAF-16 expression.

(A) ORO quantification of wild-type, ins-1(nj32) and daf-2(e1370) single mutants, daf-2(e1370); ins-1(nj32) double mutants, daf-1(e1370);Ex[ges-1p::daf-2a] lines and daf-2(e1370); ins-1(nj32); Ex[ges-1p::daf-2a] lines with (red bars) and without (gray bars) daf-16 RNAi treatment. daf-16 RNAi from L1 stage, animals shifted to restrictive temperature 25°C for 24 hours prior to assaying. (B) ORO quantification of wild-type animals treated with either EV control or RNAi targeted to daf-16 from L1 stage, shifted to HT115 + RNAi + H₂O (H₂O) or HT115 + RNAi + 40 mM glucose (Glucose) plates, containing the matching RNAi treatment 24 hours prior to staining. (C) ORO quantification of wild-type and daf-16(mu86) mutants with and without BAG-expressed ins-1cDNA (Ex[flp-17p::ins-1]) shifted to OP50 + H₂O (H₂O) or OP50 + 40 mM glucose (Glucose) plates 24 hours prior to staining. (D) Representative micrographs of DIC, DAF-16::NeonGreen (DAF-16::NG), and autofluorescence (RFP) in wild-type and ins-1(nj32) mutants shifted to OP50 + H₂O (H₂O) or OP50 + 40 mM glucose (Glucose) plates 24 hours prior to imaging. (E) Calculation of nuclear:cytosol DAF-16::NG levels in wild-type and ins-1(nj32) mutants shifted to OP50 + H₂O (H₂O) or OP50 + 40 mM glucose (Glucose) plates 24 hours prior to imaging. (F) Quantification of normalized cellular fluorescence levels (CTCF) of DAF-16::NG in the nuclei and cytosol of wild-type and ins-1(nj32) mutants shifted to OP50 + H₂O (H₂O) or OP50 + 40 mM glucose (Glucose) plates 24 hours prior to imaging. A–C and E and F significance assessed by mixed-effects model (2-way ANOVA). All data presented as $\overline{x}$ + SEM, n shown in bar. A–C and E and F significance assessed by mixed-effects model (2-way ANOVA). ns, not significant (p > 0.05), * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001. The underlying numerical data can be found in S1 Data. CTCF, calculated total fluorescence; EV, empty vector; ORO, Oil-Red O; RFP, red fluorescent protein; RNAi, RNA interference.

Fig 8. Model of the ETS-5/INS-1 regulatory module.

Under standard food conditions, ETS-5 directly promotes ins-1 expression in the BAG neurons. INS-1 acts from the BAG neurons to reduce fat levels in the intestine by activating DAF-2, which inhibits DAF-16 activity. Reduced fat stores lead to increased foraging behavior. Worms that experience excess dietary glucose down-regulate ETS-5 and INS-1 levels in the BAG neurons. This suppression of ETS-5/INS-1 leads to decreased intestinal DAF-2 activity, and increased intestinal DAF-16 levels. DAF-16 then alters intestinal gene expression to promote fat storage and produces a satiety-induced reduction in foraging activity.